Design of state predictive congestion controllers considering effective equilibrium points

Author

Azuma, Takehito ; Fukuyama, Kazuki

Author_Institution

Dept. of Electr. & Electron. Eng., Utsunomiya Univ., Utsunomiya, Japan

fYear

2010

fDate

18-21 Aug. 2010

Firstpage

520

Lastpage

523

Abstract

The purpose of this article is to design state predictive congestion controllers by considering effective equilibrium points of nonlinear models of TCP/AQM networks and to illustrate the effectiveness of the designed congestion controllers via the ns-2(Network Simulator ver.2) simulator. First nonlinear models of TCP/AQM networks derived in the paper are linearized at some equilibrium points and a method to configure the equilibrium points of the models of TCP/AQM networks is proposed by focusing on the effective probabilities of packet marks. Second congestion controllers are designed by using state predictive control. Finally the designed controllers are implemented in the router of TCP/AQM networks and some simulation results are shown to illustrate the effectiveness of the proposed method by using SIMULINK and the ns-2 simulator.

Keywords

computer networks; data communication; nonlinear control systems; predictive control; probability; telecommunication congestion control; TCP-AQM networks; effective equilibrium points; network simulator; nonlinear models; packet mark probability; state predictive congestion controller; Computational modeling; Computer networks; Equations; Mathematical model; Network topology; Predictive models; Simulation; congestion control; linear time-delay system; state predictive control;

fLanguage

English

Publisher

ieee

Conference_Titel

SICE Annual Conference 2010, Proceedings of

Conference_Location

Taipei

Print_ISBN

978-1-4244-7642-8

Type

conf

Filename

5603036